This research addresses two major issues. First is the methodological problems involved in measuring compliance, and second is the behavioral problems involved in predicting and improving compliance, in two patient populations: an indigent multi-ethnic group, and primarily Caucasian, upper middle class patients. We will use direct and indirect measures of compliance with prednisone and allopurinol among patients with newly diagnosed hematologic malignancy. Direct measures include specific blood level analysis of the drugs in question. Indirect measures include patient self report, family report, therapeutic effect, clinic appointments kept. We are particularly interested in identifying direction (over Vs. under compliance), degree, pattern, regularity, and trends (maintenance over time) in compliance. Predictors of noncompliance will be considered in five classes: cultural, social, psychological, environmental, and therapeutic. We intend to use repeated patient interviews, family interviews, and medical record reviews to acquire this data. Several standardized psychological and social adjustment scales will be integrated into the patient interviews. We will introduce three intervention packages to staggered cohorts of patients to test their effect upon compliance. The first package (Education + Re-inforcement) consists of six individual slide-tape programs, each specific to the disease, administered in inter-active format between nurse and patient. Re-inforcement is provided by a monthly medication calandar. The second package (Education + Social Support + Re-inforcement) consists of the same interventions as above; additionally, a home visit is made by nurse, to increase social support, and physically restructure the home to allow easier remembrance of pill-taking. The third package is similar to the second, except that the education component is administered in the home, to patient and family. These groups will be compared to a non-intervention control group. We will use a modified interrupted time series design, with staggered cohorts of patients phased into and out of study, allowing follow-up of each patient for six months. Relative effect of interventions in the ethnic groups will be ascertained.